Lightning Talk: Instant Backups and Restores

on YouTube ↗10m

" summary

Claim: create a cluster-wide consistent backup by taking application-level disk snapshots of coordinators TLogs and storage and restore by attaching those disk images to another cluster. Mechanism uses a snapshottable file-system or block snapshots and a low-cost snap create transaction that every TLog and storage executes at the same version so the images are point-in-time consistent. Copying disk images is far cheaper and parallelizable than logical export, enabling instant restores but requiring homogeneous clusters and snapshot-capable storage.

AI-written summary · about

" chapters (4)

1. 0:00Introduction
2. 2:13Instant Backup and Restore
3. 4:05Snapshot Create
4. 7:37Demo

" related talks

• 07mLightning Talk: Optimizing Kubernetes Multicluster Backup & RestoreHường Đinh · Open Source Summit Europe 2024
• 32mSchroedingers BackupToshaan Bharvani · OSBConf 2018
• 25mData protection design for business continuityMichael Zhang · AWS Summit ANZ 2023
• 46mLock, Stock and Backup: Data GuaranteedAndrew Moore, Jervin Real · Percona Live Open Source Database Conference 2017
• 21mSQL Server 2022 – Time to Rethink your Backup and Recovery StrategyRobert Sonders, Stephen McMaster · SQLBits 2023
• 27mBackup and Restore: Discover Best Practices for a Distributed Database Backup StrategyValentin Kulichenko, Alan Caldera · Distributed SQL Summit 2022
• 21mSQL Server 2022 – Time to Rethink your Backup and Recovery StrategyRobert Sonders, Stephen McMaster · SQLBits 2022
• 27mIt's all about the DataMichael Cade · Conf42 Cloud Native 2022